gguf better support for transformers5.0 and fix bug of Qwen3Next

[n1ck-guo]

Description

gguf better support for transformers5.0 and fix bug of Qwen3Next

Type of Change

• Bug fix
• New feature
• Documentation update
• Performance improvement
• Code refactoring
• Other (please specify):

Related Issues

Fixes or relates to #1454

Checklist Before Submitting

• My code has been tested locally.
• Documentation has been updated as needed.
• New or updated tests are included where applicable.

[Copilot]

Pull request overview

This PR aims to improve GGUF export compatibility with Transformers v5.0 and address a Qwen3Next GGUF export failure reported in #1454.

Changes:

• Re-enable GGUF-related CUDA/CPU tests for Transformers ≥ 5.0 by removing version-based skips.
• Adjust GGUF export conversion logic to handle Qwen3Next tensor handling and tweak memory-clearing behavior.
• Improve test fixture model saving by copying tokenizer.model when present, and reduce repeated calibration warnings via warning_once.

Reviewed changes

Copilot reviewed 7 out of 8 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
test/test_cuda/export/test_gguf.py	Removes Transformers v5 skip to run GGUF export tests on v5+.
test/test_cuda/advanced/test_fp8_input.py	Removes Transformers v5 skip for GGUF-related FP8 tests.
test/test_cpu/export/test_gguf_format.py	Removes Transformers v5 skip for CPU GGUF format tests.
test/helpers.py	Copies tokenizer.model into the saved tiny model directory for GGUF/tokenizer compatibility.
auto_round/export/export_to_gguf/convert.py	Adds Qwen3Next-specific tensor modification hook and changes memory clearing device selection.
auto_round/compressors/utils.py	Wraps gguf-py architecture detection to provide a clearer “upgrade gguf-py” error path.
auto_round/compressors/base.py	Switches an “insufficient samples” warning to logger.warning_once.